ERS 2024: Validation and Accuracy of the Automated AI-powered Continuous Hyfe Cough Monitoring System
Summary
The Hyfe Cough Monitor System accurately monitors coughing in problematic coughers as they go about their usual daily activities. The ability to continually and passively monitor cough has the potential to improve patient care, cough research and drug development.
The ability to passively and continuously monitor cough would significantly improve cough management and research. Recent advances in acoustic AI have prompted the development of automated cough monitoring technology. Here we describe for the first time the process and results of validating an acoustic AI based cough monitor.
We collected 20-24 hours of continuous sounds from each of 23 adult subjects while they wore a CoughMonitor and went about their usual daily activities. The watch was charged <3 ft by the bedside at night, and manual cough counting is considered the gold standard. We noted the exact time of every cough in the continuous recording using a validated annotation methodology. These results were compared to the timestamps of cough from the CoughMonitor to determine the system’s performance for each subject, for the entire group, and for subsets using event-to-event and hourly rate correlation analyses.
In 546 hours of monitoring across 23 subjects, 4,454 coughs were detected by the trained annotators. Hyfe's CoughMonitor sensitivity was 90.4% (95% CI: 89.5-91.2%), with a false positive rate of 1.03/hour (95% CI: 0.94-1.11). Hourly cough rates showed a high correlation between manual counts and CoughMonitor data (Pearson r = 0.99, OLS slope = 0.94, intercept = 0.68).
The CoughMonitor's accuracy, ease of use, and scalability suggest it could significantly enhance cough management and research.
